Tech leaders bring optimisations to deep learning in the cloud
Blog|by Leanne Bevan|4 June 2019
Microsoft and Intel have announced they are partnering to bring optimised deep learning frameworks to Azure: a new offering named ‘Intel Optimized Data Science VM for Linux (Ubuntu)’.
Deep learning is a machine learning technique that puts neural networks together with multiple layers of non-linear transformations. This enables the system to learn from data and build accurate models to help fix various machine learning issues. Examples of deep learning include: computer vision, speech recognition and language understanding. Amazingly, the technical innovations have led to human level performance in vision, speech recognition and machine translation.
Default deep learning framework builds like TensorFlow are not efficiently optimised for training and inference on the CPU, so Intel open-sourced framework optimisations for their Intel Xeon processors.
The new offer is built on top of the well-known Data Science Virtual Machine (DSVM), adding new Python environments containing Intel optimised versions of TensorFlow and MXNet. The optimisations leverage various Intel tools and libraries to enhance training and inference on the Intel Xeon processors.
When running on an Azure F72S_v2 VM instance, the optimisations bring 7.7X increased speed in training.
Find out more about the performance enhancements, additional capabilities and how to get started here.
Grey Matter is hosting a series of hands-on workshops in July with Intel Software and Microsoft experts on HPC and AI in Azure. Find out more here
Contact us to find out more about Intel and Azure. We’re an Intel Software Elite Reseller, as well as a Microsoft Gold Partner and Cloud Solution Provider (CSP) for direct and indirect sale.
Contact Grey Matter
If you have any questions or want some extra information, complete the form below and one of the team will be in touch ASAP. If you have a specific use case, please let us know and we'll help you find the right solution faster.
By submitting this form you are agreeing to our Privacy Policy and Website Terms of Use.
